First, the wiring diagram. In a bicycle computer, as a rule, a reed switch is used as a speed sensor – the simplest key that closes under the influence of a magnetic field. Further in the article, cycle computers with this type of sensor are meant. It is the magnet that is mounted on the spoke of the wheel. When it rotates, the magnet passes by the reed switch, creating a short circuit of its contacts. On an electronic counter, the pulse output has a slightly different physical basis. As a rule, it consists of two contacts coming from the secondary circuit of a transistor optocoupler built into the meter, which also serves as a galvanic isolation. In some cases, the circuit of this circuit can be supplemented with an additional transistor. With the passage of the counting pulse, this optocoupler acquires short-term semiconductivity. From all of the above, we can conclude: the pulse output of the electric meter can be connected directly to the bike computer. But, due to the semiconductivity of the optocoupler, only one of the two connection polarities will work.
Secondly, the correct setting of the bike computer for the correct display of readings. In the usual case, the bike computer setup is
in programming in entering the value of the perimeter (circumference) of the wheel, which is measured in advance or taken from the technical data for the bike. It is this parameter that serves as the scaling factor. In the case of using a bicycle computer in conjunction with an electric meter, this parameter must be calculated or selected empirically. The calculation is based on the scaling factor of the pulse frequency of the electric meter at the pulse output. This parameter is a value indicating the number of pulses per 1 kWh. It is indicated in the characteristics of the meter or on its label. Most often for single-phase home meters – 3200 pulses per 1 kWh. There are other meanings as well. For a three-phase counter, I have seen a value equal to 400. For a bike computer, the parameter “number of pulses per 1 kWh” will be equivalent to the parameter “number of wheel revolutions per 1 km”. But since in practice the wheel perimeter is usually entered, a simple recalculation needs to be done: L=1000/n. In this case, L is the required wheel perimeter, which must be entered, and n is the number of wheel revolutions per 1 km (per 1000 m), or, alternatively, the number of pulses per 1 kWh. If you enter the calculated parameter into the bike computer as the wheel perimeter, then the bike computer connected to the electric meter will show power in Watts instead of speed in km / h and consumed electricity in kWh instead of distance in km. In addition, the average and maximum speed readings (if provided by the cycling computer model) will be converted to average and maximum power consumption readings.
The meter “Mercury 201.5” with a coefficient of 3200 imp / kWh participated in my experiment. For him, L=1000/3200=0.3125 m. Such a wheel perimeter, perhaps, only for a scooter (wheel diameter is about 10 cm). Therefore, it should be assumed that not every model of a bicycle computer can provide for the input of such small values of the wheel perimeters, because in everyday life there are no bicycles with such wheels at all. But personally, I came across a Chinese SB-318 bicycle computer, in which it was possible to enter such a perimeter value (with an accuracy of thousandths). And even so, it is worth remembering that the bike computer has a minimum threshold for measuring speed, below which a bike stop and zero speed will be recorded. For the speed of a bicycle, this is not so significant, even if, for example, the minimum measured speed is 1 km / h. But in the case of power, this minimum of 1 kW in practice does not cause much interest: there is a desire to display a lower power value, at least from 100 W.
Now back to my own bike computer, which in my article is called a bike speedometer. Making a revision of the firmware, I decided to adjust it separately for using a bicycle speedometer in conjunction with an electric meter (for experiment). In addition to introducing another scaling factor (wheel perimeter), I also added the function of calculating the cost of electricity in rubles according to the current tariff. If you read carefully my article about the velospeedometer, you will notice that I specified the minimum measured speed value. It depends on the perimeter of the wheel and the maximum counting period of the MK timer before overflow. For my wheel perimeter of 2.24 meters and the maximum count of the timer is 15.1 seconds. this value will be approximately 0.53 km/h. For a “wheel” of 0.3125 meters (in conjunction with an existing electric meter) – 0.0743 km / h or kW, that is – 74.3 watts. This is already much acceptable, but in the experiment I wanted to know what minimum power the electric meter fixes. That is, I wanted to know his sensitivity, in simple terms. Therefore, I supplemented the bike speedometer firmware with virtual timers – overflow counters of the main timer. With eight such meters, the minimum measured speed (already power) turned out to be 74.3 / 8 \u003d 9.3 W, which is quite commensurate with the passport sensitivity of the electric meter. In addition, for the graph display mode, I changed the vertical scale. Initially, one cell of the display corresponds to one km / h, which is quite convenient for displaying a bicycle speed graph. But for the electricity meter experiment, I increased it 10 times: 1 display cell = 100 watts.
Both experiments with two bicycle speedometers are shown in the video. As a load, I connected a heater with two power modes. I also connected low-power loads – a 10 W light bulb and a 25 W soldering iron. But I did not film the connection data. I will only note that the meter on the light bulb did not give out a single pulse for a sufficiently long time, that is, it turned out to be insensitive to such a load.
After watching the video and comparing the two experiments, I was confused by the large difference in power readings, which is why I had to supplement this article. In the experiment with a Chinese bike computer, 2.9 kW and 1.7 kW in the second and first modes, respectively. And in the experiment with homemade – 2.54 and 1.56 kW. This fact may serve as a reason to recognize these experiments as unsuccessful. But the experiments were done at different times and under different conditions, which could affect such a difference. However, if I had only one cycle computer at my disposal, then it would not be possible to compare the results of measurements. Although, the essence of the article and the essence of the experiments is to demonstrate the technical feasibility, and not to measure the power. Nevertheless, I wanted to know why there was such a considerable difference in measurements, in my opinion.
To answer this question, I connected a third, more accurate device to the meter – a Tascam DR-05 digital audio recorder. Any recording instrument can be used to some extent as a signal recorder for further analysis and measurements. In my case, the input of the audio recorder is connected directly to the pulse output of the electric meter, and in the settings of the audio recorder, the voltage supply mode is set to the audio input to power the connected electret microphone. Instead of such a microphone – an electric meter. It is clear that it will not calculate and show the power, but it can be used to record a pulse signal for further calculations manually.
I recorded at the maximum sampling rate of 96 kHz to achieve maximum time accuracy. It is an alternation of pulses of a peculiar shape. By ear, this recording is heard as clicks, going in time with the pulses from the electric meter. Audio recording analysis consists in measuring the time interval between adjacent impulses in different sections (in Adobe Audition 1.5). The measurement data is then converted into power readings using simple calculations. That is, in this case, the audio recorder acts as a fairly accurate stopwatch. The recording was made within three minutes. The fan heater ran for approximately 1 minute at half power, 1 minute at full power, and 1 minute again at half power. In arbitrary places, I measured three intervals between pulses for one mode and three intervals for another. To achieve accuracy, the intervals were measured in the number of samples s.
In this case, the period T (sec.) is calculated as follows: T=s/96000. And power (kW) P=3600/3200/T. The calculations made in Excel are shown in the figure.
The average power indicators turned out to be approximately 2.77 and 1.43 kW for the second and first modes, respectively. Comparing these values with the values from the video with experiments, it is impossible to draw a conclusion about the correctness or incorrectness of any measurement tool. As it turned out, all three measuring instruments (a homemade bicycle speedometer, a Chinese bicycle computer and an audio recorder) measure the same within the permissible error. It turned out that the power of the fan heater floats and depends, at a minimum, on the voltage in the network, the ambient temperature and the duration of operation.